Table 5.
Mean and standard deviation of whole slide testing metrics for each model with and without augmentations.
| Model | AUG | ACC | SEN | SPE | F1 | ||||
|---|---|---|---|---|---|---|---|---|---|
| Mean | SD | Mean | SD | Mean | SD | Mean | SD | ||
| Tumor | 0 1 |
0.978 0.978 |
0.019 0.015 |
0.001 0.106 |
0.002 0.055 |
0.995
0.994 |
0.008 0.008 |
0.971 0.975 |
0.029 0.023 |
| Tumor + FIB | 0 1 |
0.979
0.978 |
0.020 0.014 |
0.001 0.241 |
0.002 0.132 |
0.991 0.992 |
0.004 0.002 |
0.972 0.977 |
0.029 0.021 |
| Tumor extended | 0 1 |
0.977 0.981 |
0.020 0.019 |
0.005
0.000 |
0.009 0.000 |
0.995 0.998 |
0.010 0.002 |
0.971 0.973 |
0.028 0.029 |
| Interhospital | 0 1 |
0.958 0.982 |
0.050 0.015 |
0.219 0.256 |
0.367 0.143 |
0.971 0.995 |
0.055 0.003 |
0.963 0.978 |
0.033 0.022 |
FIB, tumor related fibrosis; AUG, augmentations; ACC, accuracy; SEN, sensitivity; SPE, specificity; F1, weighted F1-score; SD, standard deviation. Bold values indicate the highest value of the 2 metrics per model (with and without AUG).